A Practical Significant Technic in Solving Overfitting: Regularization
Muyuan Li · Theoretical and Natural Science · 2023
The passage mainly discusses the solution to overfitting. Overfitting usually happens when people are training their machine learning models. When a model is overfitted, it only fits one particular dataset and misses most of the data points from another dataset. This problem affects the model's performance and makes it unable to use for its purpose. So how to solve this problem with significance and practical meaning? At the beginning of the passage, I will introduce some theoretical foundations for overfitting. Then I will define the concept of overfitting and show an example of overfitting in the machine learning model. After that, I will tell you how to pick the correct model with the testing set. Then, the passage focuses on the discussion of regularization, which is a helpful technique for solving overfitting. And I will compare the L1 and l2 regularization to help you find the suitable one.